mirror is brought back to its first position by the spiral hair-spring shown in the figure.

Biot found it impossible to work at his experiments on sound during the day in Paris; he was obliged to wait for the stillness of night. With the instrument just equally difficult to make ac

described I found it almost curate experiments in London. Take a single experiment in illustration. The mirror was fixed so as to cause the cross-hair of the telescope to cut the number 727 on the scale; a cab passed while I was observing-the mirror quivered, obliterating the distinctness of the figure, and the scale slid apparently through the field of view and became stationary at 694. I went upstairs for a book; a cab passed, and on my return I found the cross-hair at 686. A heavy waggon then passed, and shook the scale. down to 420. Several carriages passed subsequently, after which the figure on the scale was 350. In fact, so sensitive is the instrument that long before the sound of a cab is heard its approach is heralded by the quivering of the figures on the scale.

Various alterations which were suggested by the experiments were carried out by Mr. Becker, and the longer I worked with it the more mastery I obtained over it; but I did not work with it sufficiently long to perfect its arrangement. Some of the results, however, may be stated here.

At the beginning of a series of experiments the scale was properly fixed, and the pressure of the pointed vertical rod F, fig. 1, on the end of the iron bar, I, so regulated as to give the mirror a convenient position; then, before the bar was magnetised, the figure cut by the cross-hair of the telescope was read off. The circuit was then established, and a new number, depending on the altered length of the bar by its magnetisation, started into view. Then the circuit was interrupted, and the return of the

mirror towards its primitive position was observed. The mirror, as stated, was drawn back to its first position by the spiral hair-spring shown in fig. 1. Here are some of

Here the magnetisation of the bar produced an elongation expressed by 107 divisions of the scale, while the interruption of the circuit produced only a shrinking of 47 divisions. There was a tendency on the part of the bar, or of the mirror, to persist in the condition superinduced by the magnetism. The passing of a cab in this instance caused the scale to move from 517 to 534—that is, it made the shrinking 64 instead of 47. Tapping the bar produced the same effect.

The bar employed here was a wrought-iron square core, 1.2 inch a side and two feet long.

The following tables will sufficiently illustrate the performance of the instrument in its present condition. In each case are given the figures observed before closing, after closing, and after interrupting the circuit. Attached to each table, also, are the lengthening produced by magnetising and the shortening consequent on the interruption of the circuit :

These constitute but a small fraction of the number of experiments actually made. There are, I may add, very decided indications that the amount of elongation depends on the molecular condition of the bar. For example, a bar taken from a mass used in the manufacture of a great gun at the Mersey Iron-works suffered changes on magnetisation and demagnetisation considerably less than those recorded here.' With bars of bismuth, however strong might be the magnetism, no change whatever was observed.

1 I owe these bars to the liberality of the proprietors of the Mersey Iron-works, through the friendly intervention of Mr. Mallet.

T

6 THE POLYMAGNET.

THE polymagnet consists of an arrangement of two horseshoe electro-magnets, a helix of covered copper wire disposed between them, and suitable means of suspension.

A section of one of the electro-magnets and its surrounding spirals is given, fig. 1. ab, cd are two cylindrical cores of soft iron, which are united by a cross-piece of the same material, ef. Through the cross-piece pass the strong screws g and h into the cores, and by them the ends b and d of the cores, which are accurately planed so as to ensure perfect contact with the cross-piece, are attached to the latter. The diameter of the cores is 1.125 inch, and their distance apart, from centre to centre, 4.85 inches; the cross-piece ef is drawn in proportion.

Round each core is a helix of copper wire, overspun with cotton, saturated with shell-lac. In winding the helix, two lengths of wire, one covered with red cotton and the other with green, are laid side by side and coiled as a single wire. The diameter of the wire is 0.1 of an inch, and the weight of it which surrounds each limb of the magnet is 12 lbs. For all four limbs, therefore, a weight of 48 lbs. is made use of.

The second electro-magnet is in every respect similar to the one just described.

Fig. 2 is a front view of a flat helix of covered copper wire, intended to be placed between the two electromagnets; it has an internal diameter, ab, of 1 inch; an external diameter, cd, of 8 inches, and measures along its